Clays and Clay Minerals, 197 l, Vol. 19, pp. 3 2 1 - 3 3 3 .

Pergamon Press.

Printed in G r e a t Britain

REACTIVITY OF CLAY MINERALS WITH ACIDS A N D ALKALIES*

DOROTHY CARROLL-? and HARRY C. STARKEY

U.S. Geological Survey, Menlo Park, Calif. 94025 and Denver, Colo. 80225
(Received 16 April 1971)

Abstract--One-g samples of a montmorillonite, a metabentonite, an illite, two kaolinites, and three halloysites were treated with 50 ml of hydrochloric acid (6-45 N, 1 : 1), acetic acid (4.5 N, 1 : 3), sodium hydroxide (2.8N), sodium chloride solution (pH 6.10; Na = 35%0; C1 = 21.5% 0), and natural sea water (pH 7.85; Na = 35.5%o; CI = 21.5%o) for a 10-day period in stoppered plastic vials. The supernatant solutions were removed from the clay minerals and analyzed for SiO2, A1203, CaO, MgO, Na~20, and K20. All the solutions removed some SiO2, A1~O3,and Fe203 from the samples, but the quantities were small. Sodium hydroxide attacked the kaolin group minerals more strongly than it did montmorillonite, metabentonite, or illite. Halloysite was more strongly attacked by hydrochloric acid than was any of the other experimental minerals. Hydrochloric acid removed iron oxide coatings from soil clay minerals, but acetic acid did not remove them completely. The samples most strongly attacked by HCI and NaOH were examined by X-ray diffraction. Acid treatment did not destroy the structure of the clays, but the halloysite structure was partially destroyed. Sodium hydroxide attacked the halloysite structure, as shown by chemical analysis and X-ray diffraction. These experiments show that treatment in dilute acids has no harmful effect in the preparation of clays for X-ray diffraction. Acetic acid is preferred to hydrochloric acid for this purpose. Hydrochloric acid cleans clay minerals by removing free iron oxide from the surface; acetic acid is less effective. treatment is the solution of aluminum from the octahedral layer (Paver and Marshall, 1934). This containing clay minerals are commonly prepared aluminum has been called "exchangeable aluminiby treatment in hydrochloric, acetic, or formic num" because it is released from H-clays and is acids or by ion exchange resins in the H-form (Ray presumed to occupy the exchange positions in minerals such as montmorillonite and may be et al., 1957). Acids and ion-exchange resins are also used to prepare the H-form of clay minerals partly present as free alumina (Foster, 1953) and for studies of exchange reactions and for titration partly released from the octahedral layer. The with bases (Aldrich and Buchanan, 1958). Nutting presence of this alumina makes it difficult to use (1943) found that acids remove alkali metals, H-form clay minerals in titration studies or to alkaline earths, iron, and aluminum from clay keep H-clays for experimental purposes (Aldrich minerals and that alkalies remove silica. The first and Buchanan, 1958). This paper describes the results of treating eight effect of acids is to remove the exchangeable cations, thereby producing H-clays. The alumina clay-mineral samples with acids and alkalies for octahedral layer of a clay mineral is then attacked, 10 days at room temperature (about 25~ The and later the silica tetrahedral layer. Alkalies first acids used were hydrochloric and acetic, 6.45 N attack the tetrahedral layer. If the attack is pro- and 4.5N, respectively. The alkalies used were longed and severe the mineral structure is de- sodium hydroxide (2.8N), a 3.5% solution of stroyed, but 7 5 - 8 5 per cent of the total alumina sodium chloride (0.6 N), and natural sea water must be removed from montmorillonite before the (pH 7.85). The sea water was collected in the structure is completely destroyed (Grim, 1953, Florida Straits and was given to the authors for p. 298). The kaolinite structure resists prolonged experimental purposes by Preston E. Cloud, Jr. treatment in concentrated sulfuric acid (Murray, The salt solution was included in the experiments 1951). One of the most noticeable effects of acid because Mukherjee et al. (1947, 1948) had reported that silica, alumina, and iron were removed from the H-forms of clay minerals by leaching with a *Publication authorized by the Director, U.S. Geo- salt solution. Sea water was used so that the effects of a simple electrolyte and a polyelectrolyte could logical Survey. be compared, and also the removal of SiOz and tDeceased, January 30, 1970. 321
INTRODUCTION INSOLUBLE residues of limestones and dolomites

322

D. CARROLL and H. C. STARKEY hedral unit, inasmuch as some is present as free ferric oxide. The formula calculated is
3+ M gor 2+ [A11.56Fe0r
"

A1203 from clay minerals by sea water could be assessed. The effect of the various solutions on clay minerals was examined by chemical analysis of the solutions after they had been in contact with the minerals. X-ray diffraction studies were made of those minerals that were shown by chemical analysis to have lost considerable amounts of silica and alumina. EXPERIMENTAL CLAY MINERALS The clay minerals used in these experiments can be described structurally as in Table I. Chemical analyses of them are in Table 2.

(x = Mg, Na, Ca)33. This is almost identical with that for a montmorillonite from Upton, Wyoming, given by Foster (1951, p. 724). The total exchange for montmorillonite (A) is 87 me/100 g. The exchange capacity of the Upton montmorillonite is given as 86 me/ 100 g.

Metabentonite (B)
The metabentonite from High Bridge, Kentucky (API no. 24), consists of a mixed-layer structure of illite and montmorillonite, with illite predominating (Bradley, 1945); it has been termed a potash bentonite.

Montmorillonite (,4)
Calculation of the standard formula by the method of Ross and Hendricks (1945, pp. 41-42) indicates that the amounts of free SiO2 and free A1203 are negligible. However, Table 2 shows that some quartz is present. Some of the MgO is in the exchangeable form. It is probably not correct to place all the Fe203 given in Table 1 in the octa-

lllite (C )
The illite from Fithian, Illinois (API No. 35), is

Table 1. Structural description of experimental clay minerals Layer lattice 1 : 1, poorly crystalline 2:1 Interstratified

Mineral

1: 1

Am. Petroleum Inst. standards; API Project 49 (1949) Montmorillonite (A) Metabentonite (B) lllite (C) Kaolinite (D) Halloysite (E) x x x x

1ron oxide-coated clays separated from soils Kaolinite (F) Halloysite (G) Halloysite (H) (A) (B) (C) (D) (E) (F) x x x

Description of samples Montmorillonite, Osage, Wyoming. API std. 25b. Metabentonite, Highbridge, Kentucky. API std. 42. "lllite", Fithian, Illinois. API std. 35. Kaolinite, Bath, South Carolina. API std. 5. Haltoysite, Tintic, Utah. API std. 13. Kaolinite, Y-3, clay fraction separated from old red residual soil on colluvial material on County Road 796 near Oak Grove Church, 4.2 miles NNW of Waynesboro, Augusta County, Virginia. (G) Halloysite, 1-2, clay fraction separated from Chacha clay on Mariana limestone, about 2.5 miles southeast of Agana, Guam (depth, 6-8 in.). (H) Halloysite, 6-2, clay fraction separated from Atate clay on volcanic tuff and conglomerate, about 2.5 miles south of center of Talafofo village, Guam (depth, 2-10 in.).

REACTIVITY OF C L A Y M I N E R A L S WITH A C I D S A N D ALKALIES

323

"K

~ O ~ o o O ~ O 0- -0 O V

~
0

~
VV

~
-V

E o

O O
~D

t--, t~ "--~ ~.s ~,.~ e-, O

~ . ~ z l = ~

~:~

324

D. CARROLL and H. C. STARKEY

a mixture of 1 Md and 2M micas with montmorillonite (Yoder and Eugster, 1955, pp. 249-250). The sample used also contained small amounts of quartz, calcite, and pyrite.

Halloysite (H)
This soil clay also contains an appreciable amount of organic matter. It contains 2.6% excess SiOe, and both quartz and feldspar are present (Table 3). There is 6.6% goethite (+ hematite) and about 1.8% other impurities. This sample contains about the same amount of halloysite as (G). Samples (F), (G), and (H) are reddish-brown soil clays that are coated with adhering goethite and hematite. They are shown, by X-ray diffraction, to be poorly crystalline. These samples were included in the experiments because they are characteristic of strongly weathered soils and are the type of clay minerals that are being transported from deeply weathered land surfaces to be deposited in marine sediments. All the samples are contaminated with extraneous material. Organic matter is present in halloysites (G) and (H) from Guam and sulfur is present, presumably as pyrite, in metabentonite (B). Carbon dioxide occurs in all the samples, showing that small amounts of calcite and possibly also of dolomite are contaminants. Table 3 shows the impurities (contaminants) in the standard minerals that were estimated by chemical and X-ray analysis.

Kaolinite (D)
There is an excess of 1.1% A1203 after calculation of normative kaolinite from the analysis in Table 1. Table 2 shows that some quartz is present and that excess A1~O3 determined chemically is more than that calculated. There are about 2 per cent of constituents that are not part of the kaolinite.

Halloysite (E)
Calculation of normative halloysite from the analysis in Table 1 shows that there is an excess of 5.1% SiO2. The amount of quartz found by centrifugation (Table 3) is 6.1 per cent. There is no excess alumina, and the figure for free alumina in Table 2 apparently represents attack by acid. Other impurities in the sample amount to 0.55 per cent.

Kaolinite (F)
The impurities in this sample are 7.3% goethite, 3"0% excess SiOz, and nearly 3% other impurities. The sample, therefore, contains about 87% kaolinite. The excess SiO2 is present as quartz (Table 3), which is exceedingly fine grained and difficult to remove by centrifugation.

EXPERIMENTAL

PROCEDURE

Halloysite (G )
This sample of soil clay contains organic matter which is responsible for the low chemical total in Table 2. Calculation of normative halloysite shows that there is 0-7% excess AI2Oz, about 8% goethite (+hematite), and 2.4% other impurities. The original sample contains 89% halloysite.

Samples of the standard clay minerals, montmorillonite (A), mixed-layer mineral (B), illite (C), kaolinite (D), and halloysite (E), were ground with a pestle in a mortar to pass a 270-mesh (0-053 mm) sieve, dried in an oven at 80~ for several hours, and stored in a desiccator. Kaolinite (F) and halloysites (G) and (H) were separated from the soils in which they occurred by repeated dispersion in distilled water in an end-over-end shaker. The

Table 3. Contaminantsand exchangeable cations in experimental clay minerals Free SiO2 1.0 1.1 0.8 0.6 6-1 0.1 tr. tr. Free Free A1203 Fe203 1.7 *9-3 *4.4 1.9 *5.3 n.d. n.d. n.d. 0.2 0.5 0.6 0.1 tr. 11.0 13.4 10.8 Exchangeable cations CaO MgO Na20 K20 0.31 0.73 1.68 0.01 0.04 n.d. n.d. n.d. 0-30 0.13 0.05 tr. 0-03 n.d. n.d. n.d. 1-70 ---

Mineral Montmorillonite (A) Mentabentonite(B) lllite (C) Kaolinite (D) Halloysite (E) Kaolinite (F) Halloysite (G) Halloysite (H)

Mineral impurities identified by X-ray diffraction Feldspar, quartz Feldspar, quartz Quartz, calcite Quartz Quartz Quartz, goethite, hematite Quartz, goethite, hematite Quartz, feldspar, hematite, goethite

--n.d. n.d. n.d.

-----n.d. n.d. n.d.

*The high figure for free alumina may indicate attack of the octahedral layer by acid. Analyses (in per cent) by Dorothy Carroll, H. C. Starkey, and J. C. Hathaway, tr.-trace; n.d.-not determined; - , not present. Free SiOz = quartz.

REACTIVITY OF CLAY MINERALS WITH ACIDS AND ALKALIES silt was removed by centrifugation, and the clay fraction (< 2 ix) was dried at room temperature. Five l-g ~ of each of the eight clay minerals were weighed out in polyethylene vials (40 vials in all). Five 40- to 50-ml lots of HC1 (1:1, 6.45N), acetic acid (1:3, 4-5N), N a O H (2"8 N), NaCI (35%0, 0.6 N), and natural sea water (salinity 35.50/00, chlorinity 21-5%0, pH 7.85) were added to the appropriate vials which were then stoppered tightly and allowed to stand at room temperature (about 25~ for 10 days. The contents of each vial was mixed each day by shaking, and after 10 days the solutions were separated from the clay minerals by centrifugation. The clay minerals were washed once with distilled water, and this was added to the solution removed from each clay. Each clay mineral sample was then washed twice more with distilled water and transferred to an evaporating dish and allowed to dry at room temperature. The solutions were made up to volumes that would contain less than 25 mg of SiO2 + A1203 in a 50-ml aliquot so that they could be analyzed by the rapid methods of Shapiro and Brannock (1962) for SIO2, A1..,O3, Fe.,O:~, CaO, MgO, Nw20, and K20. In order to obtain a sharp end-point, it was necessary to precipitate the R203 group before determining MgO and CaO by versene titration. Blanks were made for each solution and figures for the blanks were deducted from those obtained by analysis of the supernatant solutions. Exchangeable cations were removed from the original clay minerals by leaching with 1 N NH4CI adjusted to pH 7 with NH4OH. Exchangeable calcium and magnesium were determined in the leachate by versene titration, and exchangeable sodium and potassium were determined by flame photometer (Table 3). The free silica present as quartz and feldspar was removed from a weighed sample of each clay by repeated centrifugation at high speed. The quantity was weighed and the material examined by X-ray diffraction (Table 3). The free alumina was determined by treating the minerals with dilute cold hydrochloric acid (0.08 N) for about 2 hr and then precipitating the AI.,O.~+ Fe203 with NH4OH. The precipitate was dried and weighed. The Fe20:~ present was determined colorimetrically with K C N S and deduced from the total weight of the precipitate to give AI203 by difference. F r e e iron oxide in samples (A), (B), (C), (D), and (E) was removed by treatment with dithionate at 40~ (Deb, 1950) and determined colorimetrically with K C N S . The free iron in samples (F), (G), and (H) given in Table 3 is from the analyses in Table 2, as all the iron in these clays is considered to be free iron oxides coating the clay particles.
RESULTS

325

Chemical analyses of supernatants

The amounts of SiO~, A1203, Fe203, CaO, MgO, Na20, and K20 dissolved from the experimental clay minerals by treatment in 50 ml of various solutions at about 25~ are given, in Table 4, as concentrations and atomic proportions in the supernatants. Table 4 shows that measurable amounts of chemical elements are dissolved from the clay minerals by the various reagents used. However, the percentages removed are very small, as shown in Table 5. The well-crystallized clay minerals, montmorillonite, metabentonite, illite, and kaolinite, are only slightly affected by the acids and alkalies at the pH values stated. The HC1 (1 : 1) removed ferric oxide impurities, magnesium and calcium (part of which may be exchangeable cations), and some silica and alumina. Sodium hydroxide removed some silica and alumina, but was less effective in removing the other constituents than HCI. Sodium chloride removed CaO and MgO, but the effect on silica and alumina was negligible. Sea water removed a small amount of silica, the largest amount being from montmorillonite, as had been previously noted (Mackenzie and Garrels, 1965, p. 58). HC1 and N a O H removed a much greater percentage of constituents from poorly crystalline clays and halloysite than from the well-crystallized clays, but H A c , NaC1, and sea water removed similar amounts from all the clays. A summary of the results of the experiments follows:

Montmorillonite (14)
The structure is that of a 2 : l-layer silicate with silica tetrahedra exposed on the outer surfaces of the finely divided platy mineral and on the upper and lower boundaries of the water interlayer. The alumina tetrahedra are sandwiched between the two silica layers, and only the edges of this layer are exposed. All the solutions removed silica, the order of removal by the reagents being N a O H >> NaCI > HC1 = H A c = sea water (Table 4). F o r A1203 the order of removal was HC1 > N a O H > H A c > NaCI > sea water. HC1 was the most effective reagent in removing Fe203. The small amounts of the remaining constituents determined were exchangeable cations (Table 3). The exposed SiO2 tetrahedra were attacked by the reagents releasing the Si originally in tetrahedral coordination. N a O H may have removed the silica layer, but the other reagents effected only a partial removal.

Metabentonite (B)
Because this mineral is a mixture of illite and montmorillonite, the results of the experiment with this mineral cannot be assessed fully. However,

326

D. C A R R O L L and H. C. S T A R K E Y

Table 4. Analyses of supernatants of the reaction of I g of clay minerals with 50 ml of various solutions for 10 days at 25~ HCI 6.45 N ppm a.p. HAc 4.5 N ppm a.p. NaOH 2.8 N ppm a.p. NaC1 Na, 35o/00 ppm a.p. Sea water Na, 35.50/00 ppm a.p.

Montmorillonite (A) SiO2 AizO3 Fe203 CaO MgO Na20 K:O 80 240 100 11 71 7 nil 1.333 2.353 0-625 0.196 1.775 0.113 -120 <20 10 24 144 116 nil 60 40 40 77 31 nil nil 2-00 580 9-666 <0-196 200 1.960 0.062 8 0.050 0.428 20 0.357 3.600 nil -1.871 n.d. --nil -Metabentonite (B) 1-000 0.392 0.25 1.375 0.775 --240 280 10 99 6 n.d. tr. Illite (C) 4-000 2.745 0.062 1-768 0'150 -tr. 200 380 16 58 36 n.d. nil 40 <20 4 119 32 n.d. nil 3.333 3.725 0.100 1.035 0.900 --0-666 <0.196 0.025 2.125 0"800 --80 <20 6 n.d. n.d. n.d. n.d. 20 20 4 n.d. n.d. n.d. n.d. 1.333 <0.196 0.037 ----0.333 0.196 0.025 -----

SiO2 A1203 Fe203 CaO MgO Na~O KeO

80 300 220 56 60 nil 2

1'333 2"941 1.375 1.000 1.500 -0.021

SiO2 AI~O3 Fe203 CaO MgO N a,.,O K20 SiO2 A1203 FezO3 SiO2 AI~O3 Fe.oO3 SiO,., AI~O3 Fe~O3

40 380 340 21 266 nil tr. 60 960 14 80 3800 n.d.* 60 860 2000

0.666 3.725 2.125 0.375 6.650 --1.000 9.412 0.087 1"333 37"255 -1-000 8'431 12.500

80 160 60 356 76 tr. nil 100 480 4 160 460 n.d.* 60 100 20

1.333 160 1-568 180 0-375 12 6-357 76 1-900 4 tr. n.d. -nil Kaolinite (D)

2.666 1-765 0-075 1.357 0.100 ---

60 140 14 104 14 n.d. nil 40 40 2 60 80 n.d.* 40 40 4

1.000 1.372 0-087 1.857 0.350 --0'666 0"392 0.012 1"000 0"784 -0"666 0'392 0.025

40 120 6 n.d. n.d. n.d. n.d. 20 80 2 20 20 n.d.* < 20 20 2

0' 666 1"176 0'037 ----0.333 0.784 0-012 0'333 0"196 -< 0'333 0.196 0.012

1.666 1120 18"666 4'706 580 5'686 0.025 nil nil Halloysite (E) 2'666 1700 28'333 4'510 1820 17"843 -n.d.* -Kaolinite (F) 1.000 2200 36"666 0'980 2000 19.608 0.125 4 0.025 Halloysite (G) 1.333 1880 31.333 1-568 1840 18"040 0.125 4 0.025 Halloysite (H) 2-000 2.745 0-025 10 1920 1720 0-166 18.823 10.750

SiOz AlzO3 Fe2Oz SiO2 A1203 Fe203

60 1480 2380 100 1340 1980

1.000 14.510 14.875 1.666 13.137 12.375

80 160 20 120 280 4

20 60 4 40 <20 6

0.333 0"588 0.025 0.666 <0-196 0.037

<20 <20 4 <20 <20 10

<0.333 <0"196 0.025 <0-333 <0.196 0-062

Analysts: Leonard Shapiro and Dorothy Carroll. p p m - p a r t s per million; a . p . - atomic proportions; t r . - trace; n . d . - n o t determined; n . d . * - n o t determined, because sample appeared colorless; - - , not applicable. Because of the limitations in the usefulness of chemical methods, N a 2 0 was not determined in the supernatants of N a O H , NaC1, or sea water, nor were CaO, MgO, Na20, K20 determined in sea water. Blanks were made for each solution used in this experiment and the figures given represent the analyses of the supernatants minus the blank determinations.

REACTIVITY OF CLAY MINERALS WITH ACIDS AND ALKALIES Table 5. Average loss (in per cent) of total constituents of experimental clay minerals Clay minerals A-D (well-crystallized) E (halloysite) F-H (poorly crystallized) HC1 6.45 N 0.08 0.39 0.34 HAc 4.5 N 0.04 0.06 0.03 NaOH 2.8 N 0.09 0.35 0.38 NaC1 Na, 350/00 0-03 0.01 0-007 Sea water Na, 35.5~ 0.01 0.004 0.004

327

the structure as a whole is less attacked by these reagents than is pure montmorillonite. The results may show that the montmorillonite component is attacked whereas the illite component is not.

lllite (C)
The illite structure seems to be less subject to attack than that of montmorillonite. As the illite structure is bonded together by K + ions, there is no interlayer space. Small amounts of SiO2 were dissolved by all the reagents. More than twice as much A1203 was removed by HCI than by N a O H , but HCI also removed iron (determined as Fe203) which was in the octahedral layer isomorphously replacing AI203.

readily removed by HCI but not by HAc. These clays are poorly crystallized and may be disordered structurally, as shown in the X-ray diffractogram of halloysite (G), Fig. 3. The poor crystallinity was confirmed by electron micrographs taken by John C. Hathaway.

X-ray examination of clay minerals

HCI and N a O H dissolved larger amounts of SiO2 and A1203 from kaolinite (D) and halloysites (E) and (G) than from the other clay minerals. X-ray diffractograms made of these three clay minerals before and after treatment are shown in Figs. 1-3. These diffractograms do not show any structural changes in these minerals after treatment in HCI; but Fig. 2 shows that in halloysite (E) there was an increase in intensity of the X-ray spacings due to quartz (this sample in its natural state contains 5-6 per cent quartz as an impurity). This indicates that some halloysite was dissolved, which increased the relative amount of quartz in the sample. The X-ray diffractogram of kaolinite (D), after its treatment in N a O H , is similar to that of the untreated sample (Fig. 1). The N a O H solution attacked halloysite (E) (Fig. 2) more strongly than did the HC1; accordingly, the amount of quartz, as evidenced by the increased peak heights, is greater. The N a O H - t r e a t m e n t of haltoysite (G) almost completely destroyed its structure (Fig. 3). Halloysite (G) and also (H) (not illustrated) have disordered structures in their natural condition. Such structures are more readily attacked than those similar to the well-crystallized structure of halloysite (E). Removal of SiO2 and A12Oz from halloysite (G) causes a relative increase of the iron oxides in the sample and therefore a high background is given with CuK~ radiation. DISCUSSION These experiments have examined four kinds of reaction of clay minerals (as 2 per cent suspensions) with various reagents that can be grouped as follows: (1) strong acids (HCI) and strong bases (NaOH); (2) weak acid (acetic acid, C H z C O O H ) ; (3) a strong salt (NaC1); and (4) average sea water (salinity 35.5% 0, chlorinity 21.5%0, p H 7.85). The suspensions of the clay minerals were

Kaolinite (D)
As kaolinite has a 1 : l-layer structure, one of its plane faces consists of silica tetrahedra and the other of alumina octahedra. The structure of wellcrystallized kaolinite is uncharged and the exchange capacity is negligible (Table 3). Kaolinite (D) contains very little iron (Table 2). In a solution the two layers are equally exposed to reaction. Table 4 shows that the silica lancer is only slightly attacked by all the reagents except N a O H . The alumina layer is attacked strongly by both acids and N a O H , but only very slightly by NaCI and sea water.

Halloysite (E)
The structure of haltoysite is similar to that of kaolinite, except that it contains interlayer water between the kaolinite-like layers. The X-ray diffractogram of the experimental halloysite (Fig. 2) shows that the sample was dehydrated and corresponds to metahalloysite in the accepted nomenclature (Brindley, 1966, p. 30). The reaction of halloysite (E) is similar to that of kaolinite (D), although the amount of both silica and alumina dissolved from this halloysite is greater for HCI and N a O H . This suggests that halloysite is not as structurally stable as kaolinite. Halloysite regains its interlayer water in sea water (Carroll and Starkey, 1960, p. 87) and expands, but this change evidently has little effect on its reactivity (Table 4). The iron-coated soil c l a y s - k a o l i n i t e (F), and halloysites (G) and ( H ) - show reactions similar to kaolinite (D) and halloysite (E). Ferric oxide is

328

D. CARROLL and H, C. STARKEY

N{~OH-TREAIE D

HCI-I'REATED

L5

,5

10

d SPACING (A.}

Fig, 1. X-ray diffractograms of kaolinite (D), untreated and after treatment in HCI and NaOH (CuK,~ radiation), reacted with the four types of solutions for 10 days in a closed system. Probably none of these reactions reached equilibrium in this time, but the reaction of sea water with clay minerals other than montmorillonite may have been about 50 per cent complete (Mackenzie and Garrels, 1965, p. 57). In the first section of the experiment (1) the clay mineral is dissolved incongruently by reaction with a strong acid (HCI) and base (NaOH). In solution the ions become completely ionized and are free to react with the silica and alumina layers of the minerals. The amounts of silica and alumina removed are given in Table 4. Kaolinite and halloysite have I : l-layer structures in which each layer would seem to be equally vulnerable to the reagents, but generally a greater amount of alumina is removed than silica and stability of the structure is attacked through the alumina layer. In montmoriUonite and illite (2 : l-layer silicates), the silica layer is more exposed than the alumina layer, but HCI removes about as much alumina as does N a O H , and the structure is primarily attacked through the alumina layer. In (2) the acetic acid used has a low concentration (1:3), and as it is a weak acid it is not completely dissociated. It has only a slight effect on the clay minerals: although an equal quantity of silica was removed from montmorillonite (A), metabentonite (B), and illite (C) as with ttCI, very little alumina was removed even from the l : l - l a y e r clays, kaolinite (D) and halloysite (E). In (3) the NaCI solution contains a higher con-

REACTIVITY OF CLAY MINERALS WITH ACIDS AND ALKALIES

~176

329

HCIT R E A T E D
Q

UNTREATED

1.5

2 d SPACING (.,~)

10

Fig. 2. X-ray diffractograms of halloysite (E), untreated and after treatment in HCI and NaOH (CuK, radiation). centration of the N a + ion than sea water (0-6 equivalent as against 0.46 equivalent), but almost the same concentration of the C I - ion (0"6 equivalent as against 0.535 equivalent). The N a + ion is completely dissociated in NaCI and 99 per cent dissociated in sea water (Garrels and Thompson, 1962, p. 57). The analyses of the NaCI supernatant may show the effect of the anion, CI-. Mukherjee et al. (1947, 1948) showed that salt solutions liberated H +, AP +, and Fe z+ ions from H-clays in an open system. Apparently the prior acid treatment removed some silica, and so alumina became available for reaction with the salt solution. Both alumina and iron may have reacted to form aluminum chloride and iron chloride, respectively. Nash and Marshall (1956, a and b) similarly showed that salt solutions would react with feldspar surfaces after acid treatment of these surfaces. Sea water (4) is a polyelectrolyte with a complex chemistry (Garrels and Thompson, 1962, p. 63). It has been observed that silica is released from certain silicate minerals by reaction with sea water (Carroll and Starkey, 1960; Mackenzie and Garrels, 1965; Siever et aL, 1965), that the dissolved silica in sea water acts as a buffer controlling the p H as well as the concentration of the chief ions in sea water (Sill6n, 1961), and that the amount of silica in sea water is controlled by the silicate minerals (Mackenzie et aL, 1967) which are added to the oceans by erosion of the land and by atmospheric winds (Griffin et aL, 1968). When detrital clay minerals and fresh water are added to oceans or are dropped by winds, the sea water gradually adjusts to its normal composition, in which it contains about l ppm SiOz. After mixing, the silicate particles are widely dispersed until they are flocculated and del:iosited on the sea floor. The silicates (clays) have been added to a closed basin, just as they were in the experiment described. In the marine environment, the N a + ion is almost completely free (Garrels and Thompson, 1962, p. 63) and it can therefore react as does the N a + ion in a solution, except that it is present with other ions and its chemical activity depends partly on the buffering action of CO~ in the sea water. In experiments with the reaction of various salt

330

D. CARROLL and H. C. STARKEY

N~OH

T R E A T E D

HCI-TREATED

U N T R E A T E D

1.5

~0

d SPaC(NG (.'~)

Fig. 3. X-ray diffractograms of halloysite (G), untreated and after treatment in HCI and NaOH (CuK~ radiation). solutions with feldspar surfaces, N ash and Marshall (t956, a and b) found that if the feldspar surface had been mildly attacked by H + ions, a zone of lattice flexibility was produced in which cation exchange occurs with cations in contact with the surface. Only the H+ ion is small enough to penetrate the feldspar lattice and thereby cause a breakdown in the normal lattice relations. Attack by the H + ion release SiOz but no AI:O~. The amount of SiO2 released from feldspar is about the same as that for quartz (solubility 6 ppm). An equilibrium may be reached between the thin zone of lattice adjustment (about 2-3 unit cells thick; Nash and Marshall, 1956b, p. 35) from which silica is removed and a salt solution or sea water independent of the pH. In sea water an equilibrium may be reached when no further chemical reaction takes place with added clay minerals: sea water ~ clay minerals. Mackenzie and Garrels (1965. p. 57) suggest that in their experiment this may take about 6 months as judged by the release of silica, although more than 50 per cent of the silica was released in 10 days from the silicates they examined. However, montmorillonite released more silica than did kaolinite, illite, muscovite, or chlorite. This difference is probably related to the different types of crystal lattice of the minerals and to the strength of the bonding of the layers. T h e layers of all these minerals, except montmorillonite, are strongly bonded, whereas the structure of montmorillonite (Fig. 4) generally shows isomorphous replacement in the octahedral layers which results in exchangeable ions in the interlayer region to preserve electrical neutrality.
CONCLUSIONS

The pH of the reagents used in these experiments ranged from less than I to over 12, thereby covering the solubility ranges for silica, alumina, and ferric oxide. During the experiments the pH of the sodium hydroxide and the p H of sea water were lowered slightly by the absorption of COx (Carroll, 1960), but the amount of CO~ absorbed by these solutions in the stoppered vials was insufficient to affect the reactions. The analyses of the supernatants are those to be expected with HCI and N a O H and 2 per cent suspensions of clay minerals, The reactions to acetic acid, though varying with the kind of cta3) mineral, are noticeably weaker than Io HC1. Acetic acid is to be preferred in preparing for X-ray examination insoluble residues containing clay such as occur in many impure limestones and other materials. The acid treatment shows that the arrangement of the silica and alumina layers in clay minerals influences the

R E A C T I V I T Y OF C L A Y M I N E R A L S WITH A C I D S A N D A L K A L I E S
9 9 . INTERLAYER..

331

c o o r d i n e t e d s i l i c o n~ . ~ .

/ymagnesium subst,tution

.7
INTERLAYER " "

k
'

II/I CHARGE DEFICIENCY

Fig. 4. Schematic montmorillonite layer structure (after Fommer, 1963). r e m o v a l of silica and alumina. T h e reactivity of both HC1 and H A c is in the order: montmorillonite (A) > metabentonite (B) > kaolinite (D) > halloysite (E) = illite (C). A d h e r i n g iron oxide prevented the attack of H A c on the o t h e r three kaolinite and halloysite group minerals (F, G , H). T h e kaolin minerals w e r e attacked more strongly by N a O H than were montmorillonite, metabentonite, or illite. Small quantities of SiO2 were r e m o v e d from all the minerals by N a C I solution and by sea water. A l u m i n a was also r e m o v e d from the clay minerals by these reagents. T h e m e c h a n i s m of re)ease of alumina is not known, but it may be a complexing of alumina with the C1- ions of the solutions if the alumina is m o v e d from the octahedral sheet to e x c h a n g e positions in the minerals, as it is with acid treatment.
A c k n o w l e d g e m e n t s - W e thank Leonard Shapiro for making many of the chemical analyses and for helpful advice in many ways, W. W. Brannock for chemical advice, critical reviews, and suggestions, and R. M. Garrels and R. O. Fournier for critical reviews.

REFERENCES

Aldrich, D. G. and Buchanan, J. R. (1958) Anomalies in techniques for preparing H-bentonites: Soil Sci. Soc. Am. Proc. 22, 281-285. Bradley, W. F. (1945) Diagnostic criteria for clay minerals: A m. Mineralogist 30, 704-713. Brindley, G. W. (1966) Discussions and recommendations concerning the nomenclature of clay minerals and related phyllosilicates: Clays and Clay Minerals 14, 27-34. Carroll, Dorothy (1960) Carbon dioxide and alumina in the potentiometric titration of H-montmorillonite: U.S. Geol. Survey Prof. Paper 4OO-B, B436-438. Carroll, Dorothy and Starkey, H. C. (1960) Effect of sea water on clay minerals: Clays and Clay Minerals 7, 80-101.

Deb, B. C. (1950) The estimation of free iron oxides in soils and clays and their removal: J. Soil Sci. 1,212220. Foster, M. D. (1951) Geochemical studies of clay minerals-l. The importance of the exchangeable magnesium and cation-exchange capacity in the study of montmorillonitic clays: Am. Mineralogist 36, 717-730. Foster, M. D. (1953) Geochemical studies of clay minerals- Ill. The determination of free silica and free alumina in montmorillonites: Geochim. et Cosmochim. Acta3, 143-154. Garrels, R. M. and Thompson, M. E. (1962) A chemical model for sea water at 25~ and one atmosphere total pressure: Am. J. Sci. 260, 57-66. Griffin, J. J., Windom, Herbert and Goldberg, E. D. (1968) The distribution of clay minerals in the World Ocean. Deep-Sea Res. 15,433-459. Grim, R. E. (1953) Clay Mineralogy. McGraw-Hill, New York, 384 pp. Mackenzie, F. T. and Garrels, R. M. (1965) SilicatesReactivity with sea water. Science 150, 57-58. Mackenzie, F. T., Garrels, R. M., Bricker, O. P. and Bickley, Frances (1967) Silica in sea water: control by silica minerals: Science 155, 1404-1405. Mikherjee, J. N., Chatterjee, B. and Banerjee, B. M. (1947) Liberation of hydrogen, aluminum, and ferric ions from hydrogen clays by neutral salts: J. Colloid Sci. 2,247-256. Mukherjee, J. N., Catterjee, B. and Ray, A. (1948) Liberation of hydrogen, aluminum, and ferric ions from pure clay minerals on repeated salt treatment and desaturation: J. Colloid Sci. 3, 437-445. Murray, H. H. (1951) The structure o f kaolinite and its relation to acid treatment. Ph.D. thesis, University of Illinois, Urbana, lllinois. Nash, V. E. and Marshall, C. E. (1956a) The surface reactions of silicate minerals-1. The reactions of feldspar surfaces with acidic solutions: University Missouri, Coll. Agr. Expt. Sta. Res. Bull. 613, 36 pp. Nash, V. E. and Marshall, C. E. (1956b) The surface reactions of silicate minerals- 1I. Reactions of feldspar surfaces with salt solutions: University Missouri, Coll. Agr. Expt. Sta. Res. Bull. 614, 36 pp.

332

D. C A R R O L L and H. C. STARKEY soils and clays: U.S. Geol. Survey Prof. Paper 205-B, 23-79. Shapiro, Leonard and Brannock, W. W. (1962) Rapid analysis of silicate, carbonate, and phosphate rocks: U.S. Geol. Survey Bull. 1144-A, 56 pp. Siever, Raymond, Beck, K. C. and Berner, R. A. (1965) Composition of interstitial waters of modern sediments: J. Geology 73, 39-73. Sill6n, L. G. (1961) The physical chemistry of sea water: in M. Sears, Ed., Oceanography. Am. Assoc. Adv. Sci. Publ. 67,549-581. Yoder, H. S., Jr. and Eugster, H. P. (1955) Synthetic and natural muscovites: Geochim. et Cosmochim. Acta 8, 225-280.

Nutting, P. G. (1943) The action of some aqueous solutions on clays of the montmorillonite group: U.S. Geol. Survey Prof. Paper 197-F, 219-235. Paver, H. and Marshall, C. E. (1934)The role of aluminum in the reactions of the clays: Chem. Indus. pp. 750-760. Pommer, A. M. (1963) Relation between dual acidity and structure of H-montmorillonite: U.S. Geol. Survey Prof. Paper 386-C, 23 pp. Ray, Satyabrata, Gault, H. R. and Dodd, C. G. (1957) The separation of clay minerals from carbonate rocks: Am. Mineralogist 42, 681-686. Ross, C. S. and Hendricks, S. B. (1945) Minerals of the montmorillonite group, their origin and relation to

R6surn6-Des 6chantillons d'un gramme d'une montmorillonite, d'une m6tabentonite, d'une illite, de deux kaolinites et de trois halloysites ont 6t6 trait6s avec 50 ml d'acide chlorhydrique (6,45N 1: 1), d'acide ac6tique (4,5N 1: 3), de soude (2,8N), d'une solution de chlorure de sodium (pH 6,10; Na = 35%o; CI = 21,5~ et d'eau de mer naturelle (pH 7,85; Na 35,5~ C1 = 21,5%o), pendant une p6riode de 10 jours dans des rifles de plastique bouch6es. Les solutions surnageantes ont 6t6 s6par6es des min6raux argileux et l'on en a d6termin6 par analyse les teneurs en SiOz, A1203, CaO, MgO, Na20, Na~O et KzO. Toutes les solutions extraient des 6chantillons une certaine quantit6 de SIO2, A1203 et Fe2Oa, mais les quantit6s sont petites. L'attaque par I'hydroxyde de sodium des min6raux du groupe du kaolin est plus intense que celle de la montmorillonite, de la m&abentonite ou de l'illite. L'halloysite a 6t6 attaqu6e par l'acide chlorhydrique plus fortement que n'importe quel autre des min6raux 6tudi6s. L'acide chlorhydrique fait disparaitre les oxydes de fer de recouvrement des min6raux argileux du sol, mais l'action de l'acide ac6tique n'est pas complete. Les 6chantillons les plus fortement attaqu6s par HCI et NaOH ont 6t6 examin6s par diffraction X. Le traitement acide ne d~truit pas la structure des argiles, mais la structure de l'halloysite a ~t~ partiellement d~truite. L'hydroxyde de sodium attaque la structure de l'halloysite comme le montrent l'analyse chimique et la diffraction X. Ces expfriences montrent que le traitement par les acides dilu6s n'a pas d'effet nuisible darts la pr6paration des argiles pour la diffraction X. Pour cet usage, on pr6f~re I'acide ac6tique ~t l'acide chlorhydrique. L'acide chlorhydrique nettoie les min~raux argileux en faisant disparaitre de la surface l'oxyde de fer libre; I'acide ac6tique est moins efficace.

Kurzreferat-Eingrammprobenvon Montmorillonit, einem Metabentonit, einem lllit, zwei Kaoliniten

und drei Halloysiten wurden mit 50 ml Salzs~iure (6,45N, 1: 1), Essigs~iure (4,5N, 1:4), Natriumhydroxyd (2,8N), Natriumchloridl6sung (pH 6,10: Na = 35%0; CI = 21,5%o) und natfirlichem Meerwasser (pH 7,85; Na = 35,5% o; CI = 21,5~ fiber einen Zeitraum von 10 Tegen in zugestopften Kunststoff-Phiolen behandelt. Die iiberstehenden L6sungen wurden von den Tonnineralen entfernt und auf SIO2, A1203, CaO, MgO, Na20 und K20 analysiert. Alle die LiJsungen entfernten etwas SiOz, AlzO3 und Fe2Oa aus den Proben, doch handelte es sich um mengenm~issig kleine Betr~ige. Natrium hydroxyd griff die Minerale der Kaolingruppe st~irker an als den Montmorillenit, Metabentonit oder lllit. Der Halleysit wurde dutch Salzs~iure st~irker angegriffen als irgendeines der anderen Versuchsminerale. Salzs~iure entfernte Eisenoxydbelage von Eodentonmineralen, jedoch wurden dieselben durch EssigsSure nicht vollst~indig entfernt. Die am st~irksten yon HC1 und NaOH angegriffenen Proben wurden durch R6ntgenbeugung untersucht. Die S~iurebehandlung zerst6rte das Gefiige der Tone nicht, w~ihrend die Halloysitstruktur teilweise zerst6rt wurds. Wie aus der chemischen Analyse und der R6ntgenbeugung ersichtlich, wurde die Halloysitstruktur durch Natriumhydroxyd angegriffen. Diese Versuche zeigen, dass eine Behandlung mit verdfinnten S~iuren keine sch~idliche Wirkung bei der Vorbereitung der Tone fiir R/Sntgenbeugung ausiibt. Essigs~iure wird r diesen Zweck gegeniiher Sa]zs~iure vorgezogen. Die Salz~ure s~iubert Tonminerale durch Entfernung des freien Eisenoxyds yon der Oberfl~iche; Essigsaure ist in dieser Hinsicht weniger wirksam. PealoMe--O6pa6orKe 50M;r pacrBopoB HCI (6,45 N; 1:1), yKCyCHO~ KI'ICYlOTbI (4,5 N; 1:3), NaOH (2,8 N), NaC1 (pH 6,10; Na = 35%0; C1 = 21,5%o) i~ MopcKo~i BO//O~ (pH 7,85; N a = 35,5%0; CI = 21,5~o) B Teqerme 10 //He,~ B 3aKpblTbIX I'IJIaCTHKOBbIX6yTbInO~a~ax 6bInI~ nO//BeprHyTbI rio 1 r. MOHTMOpI'I.rl.rlOHtITa, MeTa6eHTOHI4Ta, HJI01HTa, //ByX KaOYlttHItTOBl,I Tpex ra~yaBrJTOB. PaCTBOp Ha// oca21KOM no//Bepra.~cn aHa.rm3y Ha SiO2, AlzO3, CaO, MgO, Na20, K20. Bo Bcex pacTBopax O6Hapy~eHbI ~3Bne,~eRr~bIe H3 rYlIIHI4CTblX MHHepaYlOB SiO2, A1203 rI Fe203, HO B He6oabmnx Kona~iecTaax. NaOH BO3ZIe~ICTBOBa.rl Ha M~inepa.Tbi KaOnI, IHoao~ rpyrtrn,i c~m,rtee, tleM na MOHT-

REACTIVITY

OF CLAY

MINERALS

WITH

ACID

AND

ALKALIES

333

MOpHJIJIOHI4T, MeTa~eHTOHHT HYIH HJ/~II4T. Ha r a n n y a 3 H T H C I oKa3biBana 6onee CllJlbHOe /le~CTBI~e, qeM Ha ~ p y r n e rnHHrtCTbie MnHepa.rlbI. H C I y ~ a n a n a IlYleHKH OKHC~IOB C IIOtlBeHHblX rYlHHHCTblX M~r~epanoB, a y K c y c a a a KHCJ/OTa He oKa3blBaJIa Ha HHX HHKaKOFO BO3~Ie~ICTBH~t. O 6 p a 3 u b L HCnblTaBIUHe HaH60siee CHJIbr~oe BO3JIe~CTBHe HC1 H NaOH, 6b~aa I43yqeHbI peHTFeHOBCICHM MeTO~IOM. O K a 3 a a o c b , ~ r o KHC~OTHaa o 6 p a 6 o T K a He p a 3 p y m ~ n a cTpyKTyp~l rnHH, HO cTpyKTypa r a a n y a 3 n w a IlCrlblTaJ/a HeKOTOpOe H3MeHeHIIe. N a O n Bbl3Baa a3MeHeHtte cTpyKTypbI r a n J l y a 3 H T a KaK cJIe~yeT H3 )IaHHbIX XHMHqeCKHX aHanH3OB tl H3 peHTreHOBCKI4X ]IH~b~paKIIHOHHblX KapTHH. ~)KcHepHMeHTbl noKa3antt, qTO o 6 p a d o T K a pa36aBJIeHHbIMH KIICJIOTaMI~ He oKa3blBaeT BIOe~HOFO BYiH.qHH,qHa HpttFOTOBJIeHHe F.rltlH ]IYI~ peHTFeHOBCKOFO aHadiH3a. YKCyCHa~ KrlCflOTa 2IJI~ 3TOFO 6 0 n e e IIpHFO)IHa, ~leM c o ~ a a a a . H C I oqi4tt~aeT raHHHCTbIe MHHepaYtbI, y ~ I a ~ a c Hx noBepxHocTri OKHCabI ~<eae3a; y ~ c y c n a a K~C~OTa Me~iee 3qbqbe~Ta~aa.

CCM Vol. 19. No. 5 - E